LASA

LASA数据集是由香港中文大学（深圳）的FNii和SSE机构创建，包含10,412个高质量的CAD模型，这些模型与920个真实世界的场景扫描精确对齐。数据集的创建过程涉及专业艺术家手动制作CAD模型，确保与真实扫描的精确匹配。LASA数据集主要应用于室内场景的实例级重建和3D物体检测，旨在解决复杂场景中物体重建的难题，提高数据驱动方法的性能。

The LASA Dataset was developed by the FNii and SSE institutions at The Chinese University of Hong Kong, Shenzhen. It comprises 10,412 high-quality CAD models that are precisely aligned with 920 real-world scene scans. During its development, professional artists manually created the CAD models to guarantee precise matching with the real-world scans. Primarily utilized for instance-level reconstruction and 3D object detection in indoor scenes, the LASA Dataset aims to resolve the challenges of object reconstruction in complex scenarios and improve the performance of data-driven methods.